Introduction
Research on mental health problemsFootnote 1 following humanitarian emergenciesFootnote 2 has focused on post-traumatic stress disorder (PTSD) (Inter-Agency Standing Committee, 2007; George et al. Reference George, Kanakamma, John, Sunny, Cohen and De Silva2012). However, a broad range of mental health consequences are recognised (World Health Organization, 2007). While the prevalence rates of mental health problems during and after humanitarian emergencies tend to be higher compared with other populations, there is considerable variation (Inter-Agency Standing Committee, 2007; Steel et al. Reference Steel, Chey, Silove, Marnane, Bryant and van Ommeren2009). The average rates for depression and PTSD range between 15 and 20% (Steel et al. Reference Steel, Chey, Silove, Marnane, Bryant and van Ommeren2009). Women are especially vulnerable to common mental disorders due to lack of resources, low education, limited social support and gender-based violence (Allden et al. Reference Allden, Poole, Chantavanich, Ohmar, Aung and Mollica1996; Bhui et al. Reference Bhui, Abdi, Abdi, Pereira, Dualeh, Robertson, Sathyamoorthy and Ismail2003; Ellsberg et al. Reference Ellsberg, Jansen, Heise, Watts and Garcia-Moreno2008). A recent survey (current study site) found 38% of women were psychologically distressed, and associated with previous exposure to conflict-related trauma and major stressors such as death/disability of a significant other, financial constraints, accommodation difficulties and interpersonal problems, including marital issues (Khan et al. Reference Khan, Chiumento, Dherani, Bristow, Sikander and Rahman2015).
The evidence base regarding psychological treatments for mental health problems specifically related to trauma is mounting; however, the majority of this research has been conducted in high-income settings or through specialised professionals (Tol et al. Reference Tol, Barbui, Galappatti, Silove, Betancourt, Souza, Golaz and van Ommeren2011; Tol et al. Reference Tol, Barbui and van Ommeren2013). In order to be scalable in humanitarian emergencies, it is important to establish whether such interventions remain effective when delivered by non-specialists under supervision. Such interventions have received significant attention as part of the global mental health research agenda (Kakuma et al. Reference Kakuma, Minas, van Ginneken, Dal Poz, Desiraju, Morris, Saxena and Scheffler2011). These interventions should be brief and simple, so that they can be delivered by people in the community (Rahman et al. Reference Rahman, Malik, Sikander, Roberts and Creed2008). Additionally, interventions should address a range of outcomes (trans-diagnostic), including physical functioning and common mental health and psychosocial problems.
The WHO, as part of its mhGAP (Mental Health Gap Action Programme), has developed guidance for humanitarian settings and is testing psychological interventions delivered by non-specialists, including Problem Management Plus (PM+) (Dawson et al. Reference Dawson, Bryant, Harper, Kuowei Tay, Rahman, Schafer and van Ommeren2015; Ventevogel et al. Reference Ventevogel, van Ommeren, Schilperoord and Saxena2015). PM+ has been developed for adults affected by symptoms of common mental health problems, including depression, anxiety and those related to practical issues such as unemployment and interpersonal relationship difficulties, within contexts following exposure to an emergency.
In order to fit different service, contexts and beneficiary groups, it will be important to have both individual and group versions of PM+. This will aid accessibility by a range of target groups. The group format not only draws on the same strategies as the individual version of PM+, but also has the benefits of peer interaction and the potential for group members to be therapeutic agents to each other (Yalom & Leszcz, Reference Yalom and Leszcz2005). A meta-analysis suggests that group psychotherapeutic interventions based upon cognitive behavioural approaches are as effective as individual interventions at 6-month follow-up (Cuijpers et al. Reference Cuijpers, Van Straten, Warmerdam and Smits2008). Due to the ability to reach larger numbers of people, the group format offers a potentially scalable approach that may also be a cost-effective means of intervention.
The aim of this study was to evaluate the feasibility and acceptability of the Group PM+ intervention in rural Swat, which has recently been exposed to conflict. A key feature was to explore the feasibility of non-specialist delivery of the intervention in collaboration with the Lady Health Workers (LHWs) programmeFootnote 3.
Methods
Study design
This mixed-methods study incorporates a two arm cluster randomised controlled feasibility trial, and qualitative evaluation of the acceptability of Group PM+ intervention to a range of stakeholders.
Settings and participants
Swat is a rural district in the northern part of the Khyber Pakhtunkhwa Province, Pakistan. The majority of the population (86%) lives rural (Pakistan Bureau of Statistics, 2016). Since 2007 Swat has witnessed widespread conflict/militancy. During the active conflict between the army and militants in 2009, nearly 2.5 million people of the district were internally displaced across Pakistan (Bile & Hafeez, Reference Bile and Hafeez2009). In July 2009, these displaced persons were asked to repatriate after the Government declared the area safe, although continued military presence and sporadic conflicts between militants and security agencies continues to create social unease (Rome, Reference Rome2010).
The current trial was conducted from 21st March until 12th August, 2015 in Qambar Union Council (UC), a rural conflict-affected area in Swat. Each UC has a Basic Health Unit (BHU) staffed by a physician, a Lady Health Visitor (primary healthcare worker stationed at a health unit and providing maternal and child health services), a vaccinator, a midwife and LHWs. LHWs are trained to provide mother and child healthcare and education to a catchment area of approximately 1000 people or 150 homes, conducting monthly routine health visits (Hafeez et al. Reference Hafeez, Mohamud, Shiekh, Shah and Jooma2011). Given their existing roles, LHWs provided a link between Group PM+ facilitators/delivery agents and women in the community, overcoming a significant access barrier.
Participants were females aged 18 years or above, referred for screening based on the judgment of their LHW that they were psychologically distressed. Participants who scored above 2 on the General Health Questionnaire (GHQ) and above 16 on the WHO Disability Assessment Schedule (WHO-DAS) were invited to participate in the trial by trained assessment team. Assessment team members were females from the local community. Those Women with imminent suicide risk, cognitive impairment (e.g., severe intellectual disability or dementia) or mental disorder (psychotic disorders, substance-dependence) were excluded.
As a cluster-randomised control trial the unit of randomisation was the LHWs catchment area. Of the 24 LHWs in Qambar UC, 20 LHWs were eligible for randomisation (Fig. 1), which was carried out by an independent researcher. Based on 1 : 1 allocation ratio, 10 LHWs were randomised to the intervention and 10 to Enhanced Usual Care (EUC)Footnote 4 arm. On average six women were recruited for the trial from each LHW catchment area. When randomised into the intervention arm, the LHW along with the lay-helper informed the participant of the schedule for five consecutive weekly Group PM+ sessions, with the first session scheduled not longer than 2 weeks after the pre-intervention assessment. The two lay-helpers carried out half of the intervention sessions each.
Fig. 1. Consort flow diagram of the feasibility trial.
Intervention
Group problem management plus
PM+ is a WHO psychological intervention developed on evidence of established problem solving, counselling and behavioural techniques (Dawson et al. Reference Dawson, Bryant, Harper, Kuowei Tay, Rahman, Schafer and van Ommeren2015). Group PM+ is an adaptation of the individual intervention. The multicomponent nature of PM+ allows for many symptoms of common mental health disorders to be addressed with the one intervention (i.e. trans-diagnostic). The manual has been translated into Urdu and Pashtu, and culturally adapted to the Swat setting. This study is the first testing of Group PM+, which comprises five weekly sessions each lasting 2 h, inclusive of breaks. All LHWs in the intervention arm were provided a half-day training session by the lay-helpers regarding their roles and responsibilities in the trial. LHWs initially served to facilitate introduction of lay-helpers to community, encourage participant attendance, and provide a space in which to conduct sessions. LHWs were present when Group PM+ sessions were conducted, although they were not active facilitators of the intervention content.
Group PM+ was delivered by local female lay-helpers with 16 years of education (graduates) and with no formal training of or prior experience in mental health. Lay-helpers and participants were gender matched to enhance acceptability and ensure researcher safety (Craig et al. Reference Craig, Corden and Thornton2000). The training and supervision of lay-helpers followed an apprenticeship model (Murray et al. Reference Murray, Dorsey, Bolton, Jordans, Rahman, Bass and Verdeli2011), which involves moving the primary delivery of healthcare from specialists such as psychiatrists and psychologists to lay-helpers, and building skills through on-the-job training. Three non-specialist supervisors (HN, PA and AM) and two female lay-helpers received 6 days training by the Master Trainer (KD). Training covered knowledge of common mental disorders, basic counselling and group management skills, the Group PM+ intervention and self-care strategies. This was followed by four weeks of practice cases with weekly group supervision through Skype (2–3 h duration) by the 3 supervisors. This group supervision model was continued throughout the trial, providing opportunities for peer learning and collective problem-solving. Before delivering Group PM+ as part of the trial, lay-helpers completed competency assessments (explained below under process evaluation). These were developed specifically for PM+ and involved role-plays that were scored on fidelity to the treatment, competency of intervention delivery and counselling skills.
Enhanced usual care (EUC)
In Swat, treatment-as-usual in primary healthcare centres (PHCs) to individuals with common mental disorders usually consists of: (a) no treatment, (b) placebo-based care, as evidence-based mental health care is not currently available in PHCs, or (c) referral to the District Headquarter Hospital (DHQ) for specialised psychiatric care. For this study, care was enhanced in two ways: (a) LHWs received training in primary care referral pathways for treatment of common mental disorders; and (b) primary care physicians received the training in assessment and treatment of common mental disorders routinely taught by our partner, the WHO Collaborating Center in Rawalpindi, Pakistan. Both of these trainings were delivered by the Swat District Psychiatrist in separate half-day sessions. During the study any EUC participants with severe psychiatric disorders (e.g., psychosis) or problems (e.g., suicidality) that required immediate specialist treatment and follow-up were referred to the DHQ or BHU, depending upon their needs.
Measures
Assessments were conducted at baseline and 7 weeks after baseline assessment. The instruments have been adapted and validated for use in Pakistan. Blinding was ensured during assessments through LHW instructing participants not to disclose their allocation status. In case of disclosure the assessments were reschedule through the other assessor.
Anxiety and depression (primary outcome measure)
Psychological distress in terms of states of anxiety and depression were measured with the Hospital Anxiety and Depression Scale (Zigmond & Snaith, Reference Zigmond and Snaith1983; Mumford et al. Reference Mumford, Tareen, Bajwa, Bhatti and Karim1991), an established scale consisting of 14 items. The HADS contains two sub-scales: HADS-A (anxiety, seven items and range 0–21) and HADS-D (depression, seven items and range 0–21). Higher scores indicate more severe symptoms of anxiety and/or depression. The HADS has been translated into Urdu, and showed satisfactory reliability and validity (Mumford et al. Reference Mumford, Tareen, Bajwa, Bhatti and Karim1991).
Psychological distress (screening instrument)
This was measured by the GHQ at baseline assessment only (Goldberg & Williams, Reference Goldberg and Williams1988; Minhas & Mubbashar, Reference Minhas and Mubbashar1996). When used as a screening tool, the GHQ is usually scored bi-modally (i.e., 0–0–1–1), with scores ranging from 0 to 12. In a previous study in Pakistan, a cut-off of 3 or higher has been reported to indicate clinically significant psychological distress (Minhas & Mubbashar, Reference Minhas and Mubbashar1996).
Functioning (screening instrument)
Disability was assessed using the 12-item interviewer-administered screener version of the WHO Disability Assessment Schedule (WHODAS 2.0) (World Health Organization, 2010). WHODAS 2.0 covers six domains (cognition, mobility, self-care, getting along, life activities and participation), and assesses difficulties people have during the last 30 days. Difficulties are scored on a five-point Likert scale as none, mild, moderate, severe or extreme. WHODAS 2.0 has been validated cross-culturally and is valid as a screening and outcome measure, displaying good sensitivity to change (Andrews et al. Reference Andrews, Kemp, Sunderland, Von Korff and Ustun2009; Sousa et al. Reference Sousa, Dewey, Acosta, Jotheeswaran, Castro-Costa, Ferri, Guerra, Huang, Jacob and Pichardo2010).
Depressive disorder
The Patient Health Questionnaire (PHQ-9) is a nine-item instrument measuring presence and severity of depressive disorder (Kroenke et al. Reference Kroenke, Spitzer and Williams2001). As a severity measure, the PHQ-9 score ranges from 0 to 27, with each of the nine items scored from 0 (not at all) to 3 (nearly every day). The PHQ has been validated in Urdu (Ahmer et al. Reference Ahmer, Faruqui and Aijaz2007).
Post-traumatic stress symptoms
PTSD symptoms were measured using the PTSD Checklist for DSM-5 (Weathers et al. Reference Weathers, Litz, Keane, Palmieri, Marx and Schnurr2013). This 20-item checklist corresponds with the 20 symptoms of the DSM-5 PTSD diagnosis. Items were rated on a 0–4 scale, with a total severity score of 80. The previous version of the PCL based on DSM IV PTSD symptoms, the PCL-C, has been used previously in Pakistan (Khalily et al. Reference Khalily, Gul, Mushtaq and Jahangir2012). The PCL-5 was adapted to ask for symptoms in the last week (rather than month) to enhance sensitivity to change.
Psychological outcome profiles (PSYCHLOPS)
PSYCHLOPS consists of four questions (Ashworth et al. Reference Ashworth, Shepherd, Christey, Matthews, Wright, Parmentier, Robinson and Godfrey2004), covering three domains: problems (two questions), function (one question) and wellbeing (one question). Responses are scored on an ordinal six-point scale producing a maximum score of 18 (6 points per domain). Lower scores indicate better psychological profile. PSYCHLOPS gave participants the opportunity to identify problems they wish to address and discuss during group session. Although PSYCHLOPS has not been validated in South Asian context, the tool has shown good convergent validity, satisfactory internal consistency with measures of psychological distress, and high sensitivity to change (Ashworth et al. Reference Ashworth, Evans and Clement2009).
Adverse life events: Potentially traumatic events, day-to-day life events, and demographic information was also collected to compare the two arms for major baseline differences (Mollica et al. Reference Mollica, Caspi-Yavin, Bollini, Truong, Tor and Lavelle1992; Rahman et al. Reference Rahman, Iqbal and Harrington2003).
Process evaluation
The process evaluation involved mixed-methods process monitoring and semi-structured process evaluation interviews. Process monitoring involved analysis of supervision records and assessments of lay-helper competency. The lay-helpers were assessed by four independent assessors through application of a specially developed competency rating tool that evaluated: (a) counselling skills and (b) use of PM+ strategies with participants through direct observation of specially designed role plays. Competency was rated using a five-point Likert scale ranging from 0 (not done) to 5 (excellent).
The qualitative evaluation explored questions relating to intervention acceptability through targeted semi-structured interviews with key informants following an established approach for applied mental health research in humanitarian settings (Applied Mental Health Research Group, 2013; Green & Thorogood, Reference Green and Thorogood2014). Interviews followed a semi-structured topic guide developed for each category of key informant including intervention participants, lay-helpers, LHWs, family members of intervention participants and assessment team members. They were all interviewed through the semi structured topic guide on feasibility/acceptability of the trial procedures in the community. Interviews were documented verbatim in a written transcript for subsequent analysis.
Data analysis
Since this was a feasibility trial, no power calculations were carried out (Leon et al. Reference Leon, Davis and Kraemer2011). Feasibility testing of all the components of training, delivery and supervision requires the LHWs have a sufficient case-load. The sample size allowed each LHW to have six cases (one group), which is sufficient for a feasibility study.
We carried out intention-to-treat analysis to assess the intervention effect (participant were randomised and analysed irrespective of whether or not they received the allocated intervention as long as they completed the end point assessment). A linear mixed model was used for the analysis of continuous outcomes which had treatment as a fixed effect, baseline measurement of the primary outcome as a covariate, and cluster as random effect. The treatment difference in least squares means (treatment effect) between the two arms together with its 95% confidence interval was derived from the mixed model.
All qualitative data were analysed using thematic content analysis (Ritchie et al. Reference Ritchie, Spencer, Bryman and Burgess1994; Applied Mental Health Research Group, 2013; Green & Thorogood, Reference Green and Thorogood2014) relevant to applied research. This involved researchers familiarising themselves with interview transcripts, coding transcripts by themes and sub-themes, and finalising coding through consensus. A document containing thematic categories was then reviewed alongside interview transcripts to identify quotes to illustrate key themes. All analysis was conducted in the original language, with only the quotes presented here translated into English.
Ethical approval was obtained from the local Ethics Review Board at the Lady Reading Hospital and the WHO Ethical Review Committee. Participants were only enrolled after providing voluntary written informed consent, which included information on trial procedure, confidentiality, data management and reporting.
Results
A total of 134 participants referred by their respective LHWs were approached for screening. Fourteen participants declined participation before screening. The 120 participants who agreed to participate were screened on the two screening instruments after informed consent. After screening positive, 119 participants were invited for the trial. Of the 119, 59 participants in 10 LHW catchment areas were in the intervention arm, and 60 participants in 10 LHW catchment areas were in the control arm. Seven participants of the 119 randomised (6%) were lost to follow-up. See Fig. 1 for a detailed consort flow diagram. Table 1 shows the sample characteristics, and demonstrates the treatment arms were well balanced at baseline for demographic variables.
Table 1. Demographic variables between intervention and control group
Table 2 summarises results for the primary and secondary outcomes in the intervention and control arms on intention-to-treat analysis. On the HADS score (primary outcome measure), there was a difference in the scores of both arms with the intervention arm mean HADS score 4.73 less than the control group (95% CI: (−7.41 to −2.06, p = 0.001). This difference in scores was observed for the separate anxiety and depression scores, at 2.62 [95% CI: (−4.46 to −0.77, p = 0.008] and 2.48 [95% CI: (−4.00 to −0.97, p = 0.002], respectively. Similarly on the WHODAS (functionality), PSYCHLOPS, PTSD symptoms (PCL) and depressive disorder (PHQ) scores showed a trend in favour of intervention.
Table 2. Comparison of outcomes between control and intervention groups on intention to treat analysis
Although the study was not powered to detect clinically significant differences on outcome measures, the results indicated that PM+ had the potential to improve functioning, symptoms of anxiety and depression.
Process evaluation
Process monitoring demonstrated that participants showed good levels of intervention adherence with 36/59 (61%) attending all five sessions; and 14/59 (24%), 4/59 (6.8%), 3/59 (5%) and 1/59 completing four, three, two and one session/s, respectively. Initially, both lay-helpers achieved unsatisfactory results in their competency assessments. Additional targeted training and supervision (2 days over Skype sessions and visits) was therefore provided to support in areas where their scores were low, alongside additional practice sessions in the field under intensive supervision. Following this the lay-helpers were re-assessed and deemed competent to deliver Group PM+.
Key objectives of the qualitative interviews were to explore: The acceptability of the intervention to key stakeholders, the feasibility of LHWs supporting integration of the intervention into their routine scope of work and identifying barriers and challenges to further scale-up. Interviews were conducted with two lay-helpers, five participants who completed the sessions, four family members of the completing participants, three LHWs and two assessment team members. The number of interviews was predetermined and weighted for each category of respondent. Due to time constraints data collection did not reach saturation (no new information after data saturation).
Interviewees raised common themes which are summarised in Table 3. Overall lay-helpers, participants and their families expressed satisfaction with Group PM+, identifying its ability to address problems relating to participants’ perceived psychological well-being and daily functioning. Furthermore, respondents felt that lay-helpers developed good rapport with the participants.
Table 3. Qualitative interview results
A barrier identified by participants and family members was the duration of sessions (2 h). Furthermore, participants expectation of monetary incentives to attend sessions (in line with the routine practice of many international organisation programmes), led to difficulties motivating participants to attend, and may have influenced lower levels of intervention attendance. Finally, confidentiality issues arose in sessions containing multiple participants from the same household. Consequently, many participants felt uncomfortable disclosing their problems to the group, instead seeking to discuss these with the lay-helper separately during breaks.
Discussion
This mixed-method study assessed the acceptability and feasibility of Group PM+ through a feasibility cluster randomised controlled trial. The results show that lay-helpers can be trained to achieve the desired competency to successfully deliver the intervention in community under supervision. Overall, intervention uptake was good, with Group PM+ considered useful by participants, their families, and lay-helpers. The outcome evaluation, which was not powered, nonetheless indicated relevant improvements in symptoms of depression, anxiety; general psychological profile and functioning. Barriers identified included confidentiality issues, which suggest that sessions need to be comprised of women from different households. A recommendation is that the length of the session be reduced through shortening session breaks to minimise the time that females were away from household responsibilities.
A major strength of this study is that it has been conducted in a post-conflict area with adequate recruitment and low dropout rate. The intervention was found to be acceptable to the local community and LHWs. Furthermore, there was a high trial retention rate (94%), which further suggests community acceptability of the intervention (Ehlers et al. Reference Ehlers, Clark, Hackmann, McManus and Fennell2005; Rachman et al. Reference Rachman, Radomsky and Shafran2008). This could be due to the accessibility of the intervention in the community, which reduced the potential stigma associated with seeking mental health support (Rahman et al. Reference Rahman, Fisher, Bower, Luchters, Tran, Yasamy, Saxena and Waheed2013). Indeed, good retention was also found in a trial of group interpersonal psychotherapy in Uganda (Bolton et al. Reference Bolton, Bass, Neugebauer, Verdeli, Clougherty, Wickramaratne, Speelman, Ndogoni and Weissman2003). The feasibility of procedures and acceptability of the intervention to the community, LHWs and health system show that larger scale definitive trials are possible in the setting.
LHWs successfully facilitated the introduction of lay-helpers to the community and invited participants to the sessions. They were ideal hosts for Group PM+ as they are trusted and respected in the communities, overcoming a barrier to accessing women in need. Rural Swat may be considered a highly conservative society and without the LHWs as gatekeepers, community members are unlikely to trust and accept the research procedures and new health interventions (Mohmand & Gazdar, Reference Mohmand and Gazdar2007).
The group format of the intervention had some inherent challenges; notably that women were reluctant to discuss their problems in the presence of family members. For the definitive trial this challenge will be addressed by separating family members into different PM+ groups. In addition, the long duration of each session for the definitive trial will be managed by either reducing the duration of breaks or taking them at the end of the session, according to the preference of each group.
The lay-helpers were recruited from the local area with the aim of ensuring acceptability and an ability to relate well to their participants. There are data showing that participants can have preference for delivery agents from within the communities over outside specialists, as they are easily accessible and reduce the stigma attached to help-seeking (Singla et al. Reference Singla, Lazarus, Atif, Sikander, Bhatia, Ahmad, Nisar, Khan, Fuhr and Patel2014). Furthermore, local selection is considered to enhance the potential for future intervention scale-up. The lay-helpers were graduates with no prior experience of mental health interventions, enabling access to a large pool of potential persons to take on this role. In addition, locally embedded lay-helpers could be considered more conscious of the local context, including awareness of cultural norms and sensitivity to potential security concerns.
The supervision structures are potentially scalable, with a small number of trained specialists in another centre/city providing group-supervision to many lay-helpers through video calls. Additionally, the group format of supervision enabled lay-helper peer support and supervision, which was found to be important for managing difficult situations arising during field activities, including lay-helpers emotional wellbeing. Training of lay-helpers was followed by four weeks of practice cases with weekly group supervision, therefore to enhance transfer of skills, more emphasis on experiential components and role-plays should be considered.
In conclusion, the current study shows encouraging results regarding the feasibility and acceptability of the Group PM+ intervention delivered by lay-helpers. The intervention showed improvement in depression, anxiety and functioning, although the small sample size means the results need to be interpreted with caution, with a larger scale definitive trial required to demonstrate intervention effectiveness.
Supplementary Material
The supplementary material for this article can be found at https://doi.org/10.1017/S2045796017000336
Acknowledgements
We are grateful to all the women who participated in this study and to the LHWs. We are also thankful to the district health administration for providing logistical support.
Financial Support
The study was funded by the Office of Foreign Disaster Assistance (OFDA). The authors alone are responsible for the views expressed in this article and they do not necessarily represent the views, decisions or policies of the institutions.
Conflicts of Interest
None.
Ethical Standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, revised 2008.
Authorship
Research was conceptualised by AR & MvO. All authors have reviewed and approved the final manuscript for submission.
Availability of Data and Materials
Raw data are available as additional supporting files.
